I suspect a viable plan for this could be arranged for less than $1bn :)
I suspect a viable plan for this could be arranged for less than $1bn :)
The main reason trains aren't assigned to tracks in advance is because American passenger trains are run by incompetent people who insist that practices that are normal outside of the US are completely and totally impossible and therefore it's impossible to assign trains to tracks. There is also a subsidiary reason that the platforms at Penn Station are too narrow to have good bidirectional passenger flow, so they want to keep people off of them as much as possible.
I suspect the rationale behind not telling people beforehand is that they don't want to have to deal with people trying to get on at the same time as people leaving, but I feel like the main reason people are so aggressive about trying to get to the train as fast as possible on Amtrak is precisely _because_ the system treats them as too irresponsible to get timely accurate information. If people didn't need to worry about having only a couple short minutes to avoid missing their train and could plan ahead where to be when the train arrived, I think that a lot of people wouldn't feel the need to try to jostle through the crowd to race each other for the earlier spots in the line.
Knock on wood, but I’ve never had anyone give me so much as a sideways glance for this. The red caps will let you down early if you ask for assistance anyway.
Agreed, this is the case with the LIRR, the tracks are usually known well ahead of time and so when the train arrives, people stand to the side of the doors so anyone on the train can get off, and then relatively orderly board the train. Very little rushing involved.
- People in the northeast corridor (Boston, NYC, DC) taking Acela[0]
- People who live in rural towns without airports
The structure of Amtrak's ticket fees ensures the former subsidizes the latter, but the end result is the same: the vast majority of Americans do not give a shit about the functioning of commuter trains, so the organizations who run them are effectively vestigial.
A microcosm of this is train priority. Under US law, if Amtrak runs a train on freight rail networks, Amtrak gets priority. In reality, freight operators tell Amtrak to go fuck themselves and Amtrak duly complies. This means that their schedules have no connection to reality and trains get delayed for hours because the freight operators need to send 40,000 cars worth of coal to Texas and none of their lines are dual-tracked. Nobody at Amtrak is actually going to try forcing the issue, though, because that requires money, and running a functioning train service is Not Their Job.
Hell, I wouldn't be surprised if their inability to properly announce platforms has something to do with the priority issue, too. If you don't know when a train is going to come you can't reserve a platform for it either.
[0] Me: Mom, can we ride the Shinkansen?
Mom: We have a Shinkansen at home.
The Shinkansen at home:
But the more I think about other rail networks, the more I feel like I almost never see multi-track lines. I think these lines are also being used for freight and yet... is the US really just running uniquely long trains for freight?
Yes. By a large margin.
[1] https://www.amtrak.com/on-time-performance
[2] https://www.amtrak.com/content/dam/projects/dotcom/english/p...
The problem is that nobody in 45 of the 50 states cares about Amtrak.
If for some reason you wanted to go from the DC suburbs of Virginia to Queens a plane might be okay, but realistically planes waste so much time on security and boarding that it’s not worth it.
I take Amtrak into NY even though it's basically an hour in the wrong direction to drive to Route 128 station (a suburban station south of Boston). But I do it because I find driving into Manhattan so unpleasant in spite of probably being a bit faster.
Trains vs. planes make sense for the two ends of the Northeast Corridor especially if you're going downtown to downtown. They really don't on a price or time basis for the whole route.
It’s a bit more complex than that. All train journeys are scheduled ahead of time down to the block level, to ensure that only one train will ever occupy any given block at any given time. During this scheduling process conflicts are inevitable. Each conflict must be resolved somehow, and usually it is done by having one train wait in a siding for the other to depart the conflicted block. If one of these trains is a passenger train, then the law requires that the passenger train must always be given priority. This means that the other train must enter the siding and stop while the passenger train passes.
The freight companies always follow this rule, except for one tiny exception.
Consider what happens when one of the two trains cannot fit into any of the sidings. It wouldn’t do any good for that train to enter a siding and stop, because the back of the train would still be occupying the main line. Instead, the short train must use a siding and allow the longer one to pass.
Back in the 70s when Amtrak was founded and the law was written there just weren’t any long trains. But long trains are definitely more efficient, especially when they only have one type of cargo and all the cars are going to the same place. For one thing, the short trains need a lot more conductors and engineers than the single long train does.
So you can see that when an Amtrak train conflicts with a 2–mile–long coal train, it will definitely be the Amtrak train that ends up waiting. However, this does not _delay_ the Amtrak train, because it is built into the schedule. The Amtrak train has a published arrival time, and getting there on time assumes that it will be stopped for a certain amount of time waiting for that long train to pass.
There are people who think that the freight companies are breaking the law here, but the core fact is that the law simply doesn’t say anything about it; it simply assumes that either train could wait in the siding for the other. I believe it’s been brought up in Congress even, but there’s not been enough political will to amend the law; certainly Congress could ban long trains, or require the track owner to build sidings long enough to accommodate them, etc.
There is also the issue that a _late_ train loses priority. If an Amtrak train departs late (or is delayed during the trip somehow), then it could end up waiting on trains that it was originally scheduled to go ahead of, and this can make it later.
Funny story time: I was taking a trip on the Zephyr. We were delayed for ~20 minutes shortly after leaving Denver because the train ran over a shopping cart. The conductor got on the intercom and apologized for the delay, but of course they had to dig the remains of the cart out from under the engines and inspect them for damage. The train had been on time up til that point; they made up some time during the night, but we were still late getting in to Chicago. Fun times!
I’ll believe this when I see a real analysis, and only when I see a real analysis.
A long train is huge, and revenue per long train net of fuel seems likely to be quite high. Notably, I expect it to be dramatically higher than the cost of even the number of conductors and engineers needed to run several short trains. (Conductors and engineers are not well paid.)
The tradeoff is that a long train fails if any of the cars fails, so the actual availability of a long train falls exponentially with length. (Which is “mitigated” by what appear to be rather poor, perhaps intentionally poor, inspection practices on the parts of the rail companies, thus allowing long trains with malfunctioning cars to continue to operate.)
I would expect shorter trains to be more economical with all costs considered, especially if signaling and safety systems were upgraded to allow more efficient scheduling.
Anyway, the economics don’t really matter. The long trains exist and have to be scheduled somehow.
There are probably a few city pairs here and there like maybe Seattle and Portland. But Amtrak mostly has utility for the two halves of the Northeast Corridor where it actually has quite a bit of utility. (The whole Boston to DC run is mostly too long relative to flying.)
In the New Jersey direction you have a higher chance of encountering a train with passengers on it, because they use Sunnyside Yard in Queens for this purpose, but the yard is often at capacity or the tunnels leading to them are, so those trains may need to turn around in Penn instead.
So I think its more a keeping people off platform thing or overall zero care for customer service.
For those who have ridden LIRR on weekends, theres the old Long Beach summer route where everyone is sprinting down the stairs in flip flops carrying beach chairs too. It's truly insane.
They are very competent at keeping their jobs. Especially the guy who announced what track to go to.
Most of the layout of Penn Station is historical to its original construction (the destruction of Penn Station destroyed the hall, but the corridors below it are still there.) In the 1910s when Penn Station was built, commuting by railroad was miniscule and the future suburbs of New Jersey were still just woods and farms. So the layout gave large, wide platforms to the central tracks that hosted intercity trains, and narrow platforms to the commuter tracks that can barely host the width of a single escalator. This is a liability in 2023, because the share is overwhelmingly commuter traffic.
Through run trains also have a problem with this, because if we implemented through running, today, the platforms would not adequately clear before the next train arrived, creating crowd crush on the platforms. The station is also fully at capacity, so there isn't really a way to expand the platforms to allow through running that wouldn't cut Penn Station's capacity for the next few years, if not a decade. (To give some perspective, the Japanese do projects fairly well, and Shibuya Station reorganization/widening started in 2015 and has stuff to do until 2027.) The current plan to maybe get a start on widening the existing platforms is to demolish a block of Manhattan to build new platforms to take the load while the existing ones can go out of service. Funding to take a block of the most expensive real estate in the world is unclear.
There are a lot of platforms at Penn Station, and it seems to me that trains spend quite a long time at these platforms, in part due to issues related to passenger loading and unloaded.
What if the number of platforms was reduced? (By placing bogeys that’s are just platforms on every second or third track, for example?) This would reduce the number of usable platforms by 1/3-1/2, but it would massively increase availability space for loading and unloading passengers. If this enabled faster train turnaround, it could plausibly increase station throughput.
(Aside from being annoying and perhaps dangerous, the current scheme is absurdly inefficient. People can board a train far fast than they can go down the one or two (!) main escalators, and the train cars far from the escalators barely load at all until the nearer trains are mostly full.)
edit: I’m apparently not entirely off base. Here’s a proposal that increases throughput while reducing the number of tracks:
To give some perspective on how long a reconfiguration like this takes, JR East has been working on platform optimization like this at Shibuya Station since 2015, and the last elements will complete in 2027; and that’s at a station that started more optimally than Penn Station.
This seems to be the culture in the US as a whole.
Gun control? Completely impossible, even tho it's worked well in every first world country on the planet.
Paying employees instead of tipping? Completely impossible, would disincentivise work and the economy would collapse, even tho it works well in every first world country on the planet.
Public healthcare? Completely impossible, even tho it works well in every first world country on the planet.
I don't find it overcrowded, just inefficient. Folks sitting on steps, no other seating, etc. etc...
The announcements aren't at the last second, but they are usually after the train has final slot verification and may already be unloading. You will routinely have 5-7 min to board unless there is a schedule issue.
Perhaps in off-hours they could consider early announcement, but for high density times, the infra just isn't built for that efficiently or safely.
The old trick was to watch the green Amtrak monitors, because the NJT trains would pop up there first a few seconds before the official NJT displays. That system is gone now so you can't get an early jump on the track assignment but being in the other hallways puts you at the other end of the train from the crowd and it's easier to get your preferred seat.